Reproductive biology of the calanoid copepod, Eudiaptomus graciliodes (Lilljeborg):
Polyandry, Phenology and Life Cycle Strategies.
Abdurhman Kelil Ali
A dissertation for the degree of Philosophiae Doctor
U
NIVERSITY OFT
ROMSØFaculty of Science Departement of Ecology/Zoology
June 2007
Contents
Contents ………...……….…...……….. iii
Acknowledgements……….….…….……..iv
List of Papers………...……….……….………...…………..v
1.0 GENERAL INTRODUCTION ………...………...…1
1.1 BODY SIZE, DEVELOPMENT TIME AND LIFE CYCLE ……….…1
1.2 SEXUAL SELECTION…………...…….………...………… 5
1.2.1 MECHANISMS OF SEXUAL SELECTION…….………...………… 5
The Fisher Run-away Process………...………..…….……….. 5
Indicator Mechanisms………...………..………...………..5
Direct Phenotype benefits………...………..………..…..………….. 6
Sensory exploitation hypothesis………...…………..……..……….. 7
Sexually Antagonistic Selection………...………..….…..………..7
1.2.2 POLYANDRY ………...………..………...……….. 7
Aim of the study ………...………...…………..…..……….. 9
2.0 Study species and lakes ………...………..……..10
The lakes sampled………...………..…………..….. 10
Developmental stages………...……….……….……..10
The body and secondary sexual character..……….………..…….. 11
Spermatophore and its placement ………..……..………. 12
Female genital area ………...………..……..………..14
Mating behavior………...……….……….. 14
3.0 GENERAL MATERIAL AND METHODS …………..……….….…...……..19
4.0 RESULTS AND DISCUSSION …………..…………...…..……….. 23
4.1 Life cycle and natural mating frequency …...……….…………..……….. 23
4.2 Development time, body size and clutch size ………....…………..24
4. 3 Benefits of polyandry ………...………….….. 26
4. 4 Morphological correlates of mating status explaining multiple mating . 28 5. 0 CONCLUSION ………..….………. 31
6. 0 References ……….…………..……..…………..34
Papers Paper I ………...……..…………..….…. 43
Paper II ………...……..………. 63
Paper III ………...…….………. 81
Paper IV ………..……….………. 97
Appendixes ………...….………….…………...……. 116
Acknowledgements
First and foremost I would like to thank the Almighty God for making my academic aspiration and accomplishment a reality. Then it is my pleasure to express my heart- felt appreciation and special gratitude to Prof. Ivar Folstad for the opportunity and privilege he gave me to pursue PhD under his supervision. For that and for his unreserved support and guidance during data collection, analysis and the write up of the whole thesis, he is greatly acknowledged. He was always present when I needed his help. I would also like to thank Dr. Raul Primicerio and Dr. Ståle Liljedal for their help during data collection and analysis, for a considerable amount of helpful discussion and comments on earlier drafts of the thesis. I am grateful to Ass. Prof.
Jørgen Berge for valuable instruction and assistance on the techniques of morphological measurement and for comments on manuscripts. Thanks to Geir Rudolfsen and Lars Figenschou for helping field sample collection, and to Sissel Kaino for helping in counting copepod sample and taking body size measurements and to the Norwegian State Educational Loan Fund for granting me the scholarship to undertake graduate studies at the University of Tromsø. My special thank goes to dear Kedija Reshid for loving and understanding me, and encouraging me when I was far away from my academic atmosphere. Last but not least, I remain sincere, grateful and indebted to my mother Zubeyda Awole, my father Kelil Ali and all my sisters and brothers, whose words of encouragement, affection and prayer served me as a source of strength and inspiration throughout my study.
Tromsø, June 2007
Abdurhman Kelil Ali
List of Papers
This thesis is based on the following papers, referred to in the text by the Roman numbers given below.
Paper I
Distinct developmental phenology and reproductive strategies in Eudiaptomus graciloides populations of two neighbouring subarctic lakes.
Paper II
Morphological correlates of mating frequency and clutch size in wild caught female Eudiaptomus graciloides (Copepoda: Calanoida)
Paper III
Influence of female morphology on male mating success in the calanoid copepod, Eudiaptomus graciloides.
Paper IV
No apparent fitness benefits of polyandry in female Eudiaptomus graciloides (Copepoda: Calanoida)
1. GENERAL INTRODUCTION
1.1 BODY SIZE, DEVELOPMENT TIME AND LIFE CYCLE
Body size is an important trait affecting fitness in many species and several selective mechanisms influencing body size may be identified (Stearns, 1992; Roff, 2002). Yet, fecundity selection and sexual selection are usually the most important mechanisms (Andersson, 1994). In most ectotherms, the number of eggs a female produces increases with body size. Large females are also better equipped to thwart harassing males and reduce direct costs associated with superfluous mating (Crean & Gilburn, 1998). Male body size, on the other hand, can be important in male-male competition for females and for female choice of mates, and body size can also affect the male’s mobility and search ability (Andersson, 1994) and the ability to overcome initial resistance of females to mating (Arnqvist & Rowe, 2002; Rowe & Arnqvist, 2002).
Additionally, there is often a positive relationship between male size and success in sperm competition in many polyandrous organisms (Nylin & Gotthard, 1998).
While large body size is usually associated with greater mating and reproductive success, several factors also appear to select against it (see review Blanckenhorn, 2000). Adult body size is, for example, a function of growth rate and development time (Abrams et al., 1996), and to achieve a larger size, organisms may have to either grow faster or grow for a longer time (Stearns, 1992; Blanckenhorn, 2000; Roff 2002). However, rapid growth seems to be costly (see review: Munch & Conover, 2003; Gotthard, 2001, 2004), for example, in terms of increased predation risk at higher foraging intensities (Werner & Anholt, 1993; Abrams et al., 1996; Gotthard, 2000; Munch & Conover, 2003) or of physiological costs due to higher metabolic demands (Gotthard et al., 1994; Blanckenhorn, 1998; Fischer et al., 2004). Therefore, growth rates may not always be maximized (Gotthard, 2004). Alternatively, organisms may reach a large size at maturity by growing for a longer time, i.e., by delaying age at maturity (Stearns, 1992; Blanckenhorn, 2000; Roff 2002). However, longer development time may increase the chances of mortality before reproduction (Stearns, 1992; Blanckenhorn, 2000; Gotthard, 2001; Roff, 2002). Given high pre- reproductive mortality, organisms are predicted to decrease their age at maturity (Sibly & Calow, 1986; Abrams & Rowe, 1996), with consequent costs of reduced fecundity associated with a smaller size. Therefore, the advantages of reaching
maturity at an early age, and achieving large size may be under conflict.
Understanding the optimal balance between these conflicting demands has been the focus of research on life histories (Stearns 1992; Roff 2002). Obviously, the optimal age and size at maturity are expected to depend upon environmental conditions during development (Blanckenhorn, 1998; Nylin & Gotthard, 1998) and the period within which development is constrained (McNamara & Houston, 1996; Abrams et al, 1996;
Johansson & Rowe, 1999; Gotthard, 2001, 2004)
Environmental variables that are usually invoked to explain the change in development time and final body size include food quality and availability, temperature, and the effect of size-selective predation. For example, studied ectotherms often grow faster and become larger when food is abundant or of high quality (Atkinson & Sibly 1997; Blanckenhorn, 1998). Moreover, the majority of ectotherms grow slower but mature at a larger body size in colder environments (Atkinson 1994; Berrigan & Charnov 1994; Sibly & Atkinson 1994). Such thermal clines in body size are a common phenomenon in species distributed over broad geographic ranges, with the majority of them exhibiting larger adult size in colder environments (Partridge & French, 1996; Ashton, 2004). In addition, numerous experiments have shown that ectotherms grow to large size when reared in cooler condition (see review: Atkinson 1994, Atkinson et al., 2003). The inverse relationship between body size and temperature — referred to as the temperature-size rule — may be attributed to an evolutionary response resulting from differences among genotypes (Berven 1982; Lonsdale & Levinton 1985; Partridge et al, 1994) and/or a developmental response (phenotypic plasticity) to temperature (Atkinson 1994).
Moreover, predation, which is often size-dependent, is another important factor that affects development time and body size. For example, small predators select small prey, whereas large predators choose large prey (Nylin & Gotthard, 1998). Thus, when predation from vertebrates that prefer large prey is important, selection may favor a shorter growth period resulting in reduced age at maturity and smaller body size. As body size and clutch size are closely related, selective pressure for smaller size may also result in smaller clutch size. Additionally, in animals that carry their eggs until hatching, for example, in copepods and cladocerans, the risk of predation by visually-oriented planktivorous fishes increase with clutch size as increased clutch size increases the conspicuousness of prey (Svensson, 1995, 1997). The average
clutch size of populations exposed to such fish predation has been documented to decrease (Gliwicz, 1981; Dawidowicz & Gliwicz, 1983; Vuorinnen et al., 1983;
Gliwicz & Boavida. 1996). Consequently, smaller body size may occur as a correlated response to selection against large clutches. On the other hand, invertebrate predators preferring small prey should favour large body size at first reproduction (Lynch, 1980a, b; Paine 1965; Bertness 1981; Berven & Gill 1983). The different factors favouring or opposing large body size could vary spatially and temporally; under such conditions, discrete populations should evolve different body sizes to suit local selective regimes. Therefore, it is essential to assess environmental conditions during development in order to make reliable predictions about optimal growth trajectories and final body size.
It has also become evident that the seasonal environments in which most organisms live, impose a time constraint on life histories (e.g., Forrest 1987, Nylin et al., 1989, Rowe & Ludwig 1991, Wiklund et al., 1991, Rowe et al., 1994, Abrams et al.,1996).
Especially organisms living in temperate and arctic regions experience only a limited period each year in which conditions for growth and reproduction are optimal. The life cycles of organisms under such condition must be completed within the limits set by seasonality and their fitness is often strongly dependent on successful timing of certain critical life cycle events (phenology) in relation to seasonal changes in biotic and abiotic factors (Gotthard, 2001, 2004). For example, for successful foraging in natural environments, it is important to synchronize the stages that are involved in growth and development with the periods of high abundance and quality of food;
otherwise, organisms will experience a strong reduction in fitness in terms of reduced survival or fecundity. Moreover, in order to survive the winter conditions, most temperate organisms have to reach a certain ontogenetic stage, capable of over- wintering. For example, at high latititudes, where the winter conditions are very pronounced, copepods have developed two principal forms of overwintering strategies: the adoption of a resting (or diapause) stage within the developmental sequence together with an ontogenetic, seasonal vertical migration or, alternatively, the production of resting (or diapausing) eggs (Muchline, 1998). Additionally, in sexually reproducing species mating and reproductive activities must start at a time of the year when sexually mature mates are likely to be present (Gotthard, 2001, 2004).
Therefore, organisms in seasonal environments are likely to experience selection, not
only for utilizing the available time efficiently, but also for synchronization of the life cycle. Under such condition, life history theory suggests that the optimal development time and final body size should vary with time horizons for juvenile growth and development (Werner & Anholt, 1993; Abrams et al., 1996). Empirical studies comparing the timing and duration of critical life cycle events of populations living in different biotopes may therefore lead to a better understanding of the mechanisms linking life history traits and strategies to local environmental conditions.
1.2 SEXUAL SELECTION
Darwin proposed the theory of sexual selection to account for major problems for his theory of evolution by natural selection: why have males of many animals evolved extravagant secondary sexual characters that probably reduce survival? Why are males and females so different in appearance in many species? Sexual selection, Darwin's key explanations, was his shorthand phrase for selection that arises through competition over mates (Andersson, 1994). He posited that sexual selection is different from natural selection. He wrote:
"sexual selection … depends not on a struggle for existence in relation to other organic beings or to external conditions, but on a struggle between individuals of one sex, generally the males, for the possession of the other sex. The result is not death to the unsuccessful competitor, but few or no offspring" (Darwin, 1859, p88)
Sexual selection could occur either through mate contests, usually male–male competition for mating opportunities, or as a consequence of mate choice, usually female choice for attractive males (Darwin 1871). The fundamental reason for such divergent mating tactic between the two sexes is their difference in gamete size (anisogamy) and associated investment in offspring production. As females generally produce large nutrient-rich eggs and often care for the young, they commonly invest more per offspring than do males. Consequently, males are expected to increase their reproductive success by mating with as many females as possible and, where mating opportunities are limited, to compete for mates. Females, in contrast, are expected to maximize their reproductive success by choosing a high quality male (Bateman 1948;
Travers, 1972; Clutton-Brock and Parker 1992).
1.2.1 MECHANISMS OF SEXUAL SELECTION
Darwin’s process of sexual selection, mate contest and mate choice, has long been accepted, and the ideas still represent the major concepts of modern sexual selection theory (Andersson and Iwasa, 1996; Kokko et al., 2006). However, the mechanisms by which sexual selection operates are still heavily debated. The controversy over this issue is mainly about what drives female mate choice for male traits. Currently, there are different hypotheses about the mechanisms of sexual selection in general and the evolution of female preference for male traits in particular (Andersson, 1994).
The Fisher Run-away Process
After Darwin, the next major development in sexual selection theory came in 1930 by Ronald Fisher. He filled the gap left by Darwin by proposing a verbal model of female choice. According to this process an initial female preference for arbitrary male character could reach a threshold frequency in the population that selects for more preferred males. Females that mate with males expressing the preferred traits produce more attractive sons who also carry the genes for the mating preference for more preferred male traits. As a result, a genetic coupling (Lande, 1981) will establish between the female preference and the preferred male trait, which in turn will lead to a positive feedback. This creates a self-reinforcing system that continually will exaggerate the female preference and the male trait in the population until the size of male trait compromises male survival to the extent that extra mating advantage cannot balance the reduced survival (Lande, 1981; Kirkpatrick, 1982). However, in spite of being plausible, this process have been criticized for failing to explain how preference could generally be maintained in the face of the costs of choice that might be involved (Andersson, 1994).
Indicator Mechanisms
Fisher (1915) also provided the first outline to another influential idea, variously termed as ‘‘viability-indicator’’, ‘‘handicap’’ or ‘‘good genes’’ model, that sexually selected traits in males could serve as indicators of viability for female choice. This idea received even less attention than his other work on sexual selection, and was largely forgotten (Andersson, 1994). However, the 1960s and 1970s witnessed a revival of interest in Fisher’s (1915) indicator idea. The same idea was re-introduced
and became more generally known when discussed first by Williams (1966) and later by others (Zahavi, 1975, 1977; Hamilton & Zuk, 1982). An idea much debated under this model was the handicap principle proposed by Zahavi (1975). He presented a verbal idea as a strict alternative to Fisherian runaway selection. Zahavi (1975) emphasized that indicator traits must reduce viability, that is, handicap the survival of the male and the ability to express the traits provides a test of male’s genetic quality.
Such costly traits should enforce honest signalling by males and thus serve as reliable mate-choice cues for females. Consequently, females enhance the overall quality of genes that they pass on to offspring by choosing the right male (Zahavi, 1975). Co- evolution between the female preference and the preferred male trait would then occur, resulting in exaggeration of display traits with sizable survival costs to its bearer. Several theoretical studies have later provided support for the importance of indicators processes and for the evolution of costly male traits and, also, costly female preference (e.g. Iwasa, et al., 1991; Iwasa & Pomiankowski, 1994; 1999;
Pomiankowski & Møller, 1995; Rowe & Houle, 1996; Houle & Kondrashov, 2002)
Direct Phenotype benefits
The above two models of sexual selection both predict indirect benefits to females that exert mate choice. Direct selection for female mate choice could instead arise whenever preferences directly affect female fitness. For example, males may vary in the amount of food they provide, the quality of the nest site they hold, or some other resources that directly affect female fitness. Selection favours females that can recognize males that are superior providers by some feature that is correlated with their ability to provide (see review: Kirkpatrick and Ryan, 1991; Andersson, 1994).
Using genetic models, Heywood (1989) and Hoelzer (1989) suggested that non- heritable variation in parental ability may lead to the evolution of male traits that advertise high parental ability. Unlike indirect genetic benefit models, such direct benefit model does not require mechanisms that maintain genetic variance (Andersson, 1994). Additionally, it is easy to see how male traits and female preference for honest signals of a male’s ability to increase her fitness can evolve (Andersson, 1994). There is much empirical evidence supporting the direct benefit model and more than for the indirect genetic benefit models. Yet, the latter models may be harder to evaluate, demanding a more sophisticated experimental approach.
Sensory exploitation hypothesis
The sensory exploitation hypothesis (WestEberhard, 1984; Ryan, 1990) is a rather different explanation for the evolution of female preference and male traits.
According to this model, preferences may be maintained as pleiotropic effects of natural selection on female sensory systems in the contexts of other selection regimes than mate choice, such as foraging or predator evasion. The male trait is favored simply by virtue of its manipulative effect on a pre-existing bias in the sensory system of females. Contrary to the Fisherian run-away process and indicator mechanisms, the male trait and female preference for the trait are not expected to have a tightly correlated evolution.
Sexually Antagonistic Selection
Most recently it has been suggested that sexual conflict between the two sexes over various aspects of reproduction (e.g., mating rate, female proximate fecundity) could result in the evolution of female mate choice (Arnqvist, 1992; Rowe et al., 1994;
Hollan & Race, 1998; Gavrilets et al 2001; Chapman, et al., 2003). At the centre of this conflict is the hugely different investment in the gametes by males and females (anisogamy; Bateman, 1948). Since females generally invest considerably more than males, males benefit more from increasing the number of mates and the number of matings than do females (Bateman, 1948; Parker 1979). Thus, males are expected to attempt to seduce or coerce females to mate at a rate beyond their optimum. As mating sub-optimally is costly to females, this generates counter-selection on females to evolve resistance, rather than preference, to the male trait, followed by selection favouring a more extreme male trait to overcome female resistance. Consequently, this will result in the evolution of sexual selected characters along antagonistic, rather than mutualistic, trajectories, which may ultimately lead to viability selection checking further elaboration of the male trait in question (Hollan & Race, 1998;
Gavrilets et al., 2001; Chapman et al., 2003; Arnqvist & Rowe, 2005).
1.2.2 POLYANDRY
During the last decade, the increasing use of molecular techniques has revealed that females of many species, spanning a broad array of taxa, mate with multiple males during one reproductive event (polyandry) (Birkhead & Møller, 1998; Arnqvist and Nilsson 2000; Jennions and Petrie 2000). This challenges the long held view of
females as the choosy, essentially monogamous sex (see above). Additionally, given that mating is a costly activity in terms of energy or foraging time (Wilcox, 1984;
Bailey et al., 1993; Clutton-Brock & Langley, 1997; Watson et al., 1998), predation risk (Arnqvist, 1989; Fairbairn, 1993; Magnhagen, 1991; Rowe, 1994; Ward, 1986) and disease transmission (Daly, 1978; Hurst et al., 1995), polyandry raises the question of why females commonly mate with multiple males (Andersson &
Simmons, 2006).
A number of hypotheses have been proposed to explain the evolution and maintenance of polyandry in animal mating system that in some ways parallel the mechanisms of mate choice evolution (see above) (Andersson & Simmons, 2006).
These hypotheses can broadly be categorized in to those that based on direct (material) benefits, or indirect (genetic) benefits to female following multiple mating (Hosken & Stockley, 2003). In the former case, polyandrous females may derive benefits such as nutrient received from spermatophores or ceminal fluid, an adequate sperm supply/fertility insurance or additional parental investment that cause female reproductive rate to increase with the number of mating (Thornhill & Alcock, 1983;
Arnqvist & Nilsson, 2000; Wiklund et al., 2001; Hosken & Stockley, 2003).
Additionally, females may gain indirect genetic benefits that increase offspring fitness (Jennions & Petrie, 2000). Such indirect benefits might arise from improving on earlier mating (good genes models; Thornhill & Alcock, 1983; Simmons, 1987;
Olsson, et al., 1996), reducing the risk of genetic incompatibility (Zeh & Zeh, 1996, 1997), inbreeding avoidance (Tregenza & Wedell, 2002), and increased genetic diversity (Baer & Schmid-Hemel, 1999). The different benefits hypotheses have been empirically documented for a variety of organisms (see review: Arnqvist & Nilsson, 2000; Jennions & Petrie, 2000; Hosken & Stockley, 2003). However, the extent to which these hypotheses might apply across taxa is challenged because there are increasing number of cases in which benefits to females could not be found (Kolodziejczyk & Radwan, 2003; Brown, et al., 2004; Maklakov & Lubin, 2004;
Patricia & Graeme, 2004)
An alternative hypothesis based on sexual conflict has been proposed to addresses the occurrences of polyandry. This hypothesis invokes that polyandry may evolve as a result of selection on males to copulate as often and with as many females as possible
(e.g., Parker, 1979; Rowe et al., 1994; Holland & Rice, 1998; Arnqvist & Nilsson 2000). As males benefit more from higher mating rate than do females (Bateman, 1948), they may seduce or coerce females to mate at a rate beyond their optimum, while females are expected to resist mating attempt to reduce direct costs associated with superfluous mating (Parker, 1979; Arnqvist & Nilsson 2000; Gavrilets et al., 2001). However, pre-mating struggle/resistance is also a costly behaviour to females in terms of energy (Watson et al., 1998), foraging time (Wilcox, 1984) and predation risk (Rowe, 1994) and in the face of repeated male harassment, these costs may be even more than the cost of mating itself (see above). Under such condition, female might be better off by accepting superfluous mating, in a strategy of ‘making the best of the bad job’. This phenomenon - termed as ‘convenience polyandry’ - has been proposed as an explanation for the occurrence of multiple mating in many organisms (Thornhill & Alcock 1983; Arnqvist, 1989, 1992; Rowe, 1992; Rowe et al., 1994;
Watson et al., 1998; Weigensberg & Fairbairn, 1996
Aim of the study
The main objective of this project was to address the question of multiple mating in females of the model organism Eudiaptmus graciloides (Lilljeborg). As background knowledge, I also investigated phenology and reproductive strategies of populations living in two neighboring lakes, Takvatn and Fjellfrøsvatn. The specific objectives were:
1. to assess and compare timing of certain critical life cycle events of the two populations in relation to seasonal changes (paper I).
2. to assess and compare development time, body size, clutch size of the populations in relation to environmental conditions during development (paper I).
3. to determine the natural female mating frequency (paper I).
4. to investigate the influence of female morphological traits on the opportunity for multiple matings in wild caught females (paper II, paper IV).
5. to assess the relationships of female’s morphology with egg number and egg size (paper II).
6. to investigate the influence of female morphological traits on the outcome of mating encounter under laboratory condition (paper III).
7. to examine direct fitness benefits of multiple mating to female (paper IV).
2.0 Study species and lakes The lakes sampled
Copepods are small aquatic crustaceans. They are among the most numerous multicellular organisms on earth (Mauchline, 1998) constituting the major components of most marine and fresh water zooplankton communities. I have studied the freshwater calanoid copepod species Eudiaptomus graciloides (Lilljeborg) throughout this thesis. Studies on sexual selection included in this thesis were carried out on the population from Lake Takvatn (paper II, III & IV), which showed a greater variation in life history traits and mating frequencies. Comparative work between lakes on life history traits and reproductive strategies of E. graciloides also included material from lake Fjellfrøsvatn (paper I). The two sub-arctic lakes (see map: Fig. 1) are situated in northern Norway (69°07¢N, 19°05E¢). They are characterized by short productive season and cool, oligotrophic waters with similar physico-chemistry and similar ice and snow conditions (ice thickness = 20 – 100 cm & snow depth 5 – 100 cm) (Dahl-Hansen et al., 1994; Primicerio, 2000; Klemetsen et al., 2003). Their surface areas are 15 and 6.5 Km2 respectively, and both have a maximum depth of 80 m. The fish community is dominated by Arctic char (Salvelinus alpinus), which has been introduced into Takvatn from Fjellfrøsvatn (Svenning & Grotnes, 1991). Unlike Takvatn, Fjellfrøsvatn have two morphs of Arctic char, differing in size and timing of reproduction (Klemetsen et al., 2003). Together with the small morph, Planktivorous char is more abundant in Fjellfrøsvatn, implying a greater predation risk for Eudiaptomus in this lake (Dahl-Hansen et al., 1994; Klemetsen et al., 2003; Knudsen et al., 2007). For example, Klemetsen and colleague (2003) found that food consumption of Arctic char in Fjellfrøsvatn, without including the small morph, were consistently about twice as high as in Takvatn during winter. Arctic charr are known to feed under the ice (Klemetsen & Grotnes, 1975), at temperatures close to zero (Brännäs & Wiklund, 1992). It has a retinal flexibility (Ali et al., 1984) that allows feeding at low light intensities despite the weak light of the polar night.
Developmental stages
Diaptomid copepods have complete metamorphosis. The eggs hatch into larva called Nauplii. There are six naupliar stages, usually abbreviated N1 to N6. N1 has no oral
apparatus, lives on its vitelline reserves and barely move. This suggests that the amount of reserve, which is a function of N1 size, can have an important implication on the survival of the non-feeding stage (Cooney & Gehrs, 1980; Wyngaard, 1986;
Guisande et al., 1996). After moulting in to N2, it starts to feed. N6 moults to the first of the five copepodid stages, abbreviated to C1- C5 and finally C5 moults into adults (Dussart & Defaye, 1995; Muchline, 1998).
Figure 1. Map showing the situation of the two examined lakes in Balsfjord, Troms, Northern Norway. Takvatn (214 m.a.s.l.) and Fjellfrøsvatn (125 m.a.s.l.) have a respective surface area of 15 Km2 and 6.5 Km2. Water in rivers from the two lakes merges before passing Skjold. Yellow and white lines are roads.
The body and secondary sexual characters
The body of an adult copepod is divided in to three main regions: the cephalosome, metasome and urosome (paper II, Fig. 1). The first two regions are clearly defined parts of the body and they are together known as the prosome, or anterior part of the body. The length of the prosome is usually used as a direct measure of body length or size (Dussart & Defaye, 1995; Muchline, 1998). The urosome, or posterior part of the body, consists of the genital somite and several segments posterior to it, including a pair of furca. Sexually dimorphic characters that develop during the later copepod stages distinguish females and males. Males are usually smaller in size than females
(mean prosome length in mm ± SD: male = 0.773 ± 0.018 and female = 0.857 ± 0.028; n=39) and have additional segments in the urosome (appendix 1A, 1C, 1D &
1I). The right antennule of males in calanoid (and both antennules in cyclopoids and harpacticoids) copepods is structurally modified (appendix 1A, 1B) for grasping the female preceding mating (Blades & Youngbluth, 1980; Gilbert & Williamson, 1983;
Dussart & Defaye, 1995; Muchline, 1998). Its mid-part is enlarged consisting of a hinge that enables the antennule to fold back upon itself (appendix 1B). Moreover, the fifth pair of swimming legs of male calanoid (and harpacticoids but not cyclopoids) is sexually dimorphic, asymmetrical and adapted for mating activities (appendix 1E, 1F). Its right exopod is modified into a large chela that is used to grasp the female urosome during copulation (appendix 1F). The short and stubby external spine of the left exopod is modified to hold the spermatophore and attach it on the female gential area (appendix 1E). In females, the fifth pair of swimming legs is only slightly modified (appendix 1G).
Spermatophore and its placement
Copepods reproduce sexually requiring copulation. During copulation, the male transfer spermatozoa and associated seminal secretions contained in a package called spermatophore to the female (Fig. 3 & 4). Male E. graciloides can attach only one spermatophore in one mating event and the average time between the production of two spermatophores is 43 hours (SD ± 27.1, n = 18; 3 ± 1°C; 18 : 6 h light : dark cycle). The spermatozoa of calanoid copepods are aflagellated and immobile. The development of spermatophores provides an efficient mechanism to transfer gametes to the females (Blades-Eckelbarger, 1991) and it might represent an adaptation to such limitations. Spermatophores of E. graciloides (and from the majority of other calanoid copepods) are simple, tube-shaped flasks with mean length of 326 µm (SD:
± 14, n=100) that narrows into a spermatophore neck (Fig (4A) towards its open end (Blades-Eckelbarger, 1991; Hosfeld, 1994). Such simple spermatophores usually adheres to the female by means of cement like substance present on the outside of the spermatophore neck or by secretion extruded from the spermatophore itself (Blades- Eckelbarger, 1991; Hosfeld, 1994; Defaye et al., 2000). However, in some calanoids, the spermatophores possess complex chitin-like plates, termed coupling plates by which they are attached by the male on the female urosome (Blades & Youngbluth, 1979, 1980; Blades-Eckelbarger, 1991). Full spermatophores (Fig. 4A) appear in light
microscope to contain two distinct kinds of materials, one less dense, occupying most of the spermatophore, the other, much denser, located close to the neck (Defaye et al., 2000). A completely discharged spermatophore (Figs. 5 & appendix 2C) contains nothing but some remaining secretions, which looks foamy and blistered on higher magnification (Hosfeld, 1994). Placement position of spermatophores on the female urosome differs between species (see: Muchline, 1998; Defaye et al., 2000). In many species of calanoid copepods, males attach the spermatophore over the genital pore within the genital field of the genital double somite, referred to as the direct/correct placement. In other species, however, spermatophores can be indirectly/alternately placed on parts of the genital somite remote from the genital field. Some of these indirectly/alternately placed spermatophores develop a fertilization tube, a prolongation of the original, short spermatophore neck (Hosfeld, 1994). The fertilization tube establishes connection to the genital opening. In yet other species both direct and indirect placements of spermatophores have been observed.
(Muchline, 1998; Defaye et al., 2000). In E. graciloides, the main spermatophore placement position might be direct. I never observed spermatophores attached far from the genital field. (Figs 3, 5, 6, 7 & appendix 2A). Even females with multiple spermatophores had spermatophores located closely concentrated around the genital field (Appendix 2B). In some cases, single and multiple spermatophores with fertilization tubes were observed in the direct placement position (Fig. 5, 6 & 7 Appendix 2B & 2C).
The phenomenon of multiple spermatophores has been considered a waste of reproductive effort by males (e.g. Katona, 1975; Blades, 1977; Hopkins & Machin, 1977; Hopkins, 1982; Swenson, 1997). Furthermore, the first attached spermatophore on the genital opening of female has been suggested to act as a mechanical barrier that effectively prevents insemination (discharge of seminal products) and subsequent fertilization by other spermatophores (Cuoc et al., 1989a, b; Barthélémy et al., 1998;
Defaye et al., 2000). I, on the other hand, found multiple spermatophores with fertilization tubes connecting directly to the genital opening of female (Fig. 6).
Additionally, electron microscopy studies revealed that multiple spermatophores that was attached on the female genital area reaching the female genital atrium, where fertilization of ova occurs (see appendix 3). Moreover, micro-satellite analysis has recently confirmed the occurrence of multiple paternities within clutches produced by
females that had several attached spermatophores (Todd et al., 2005). In sum, these latter observations clearly suggest that attachment of multiple spermatophores on females may not be a waste of reproductive investment.
Female genital area
The genital area of females, where males attach their spermatophores, are located on the ventral face of the genital double somite (Fig. 2). In external view, it is covered with a plate-like flap known as operculum, which is visible in non-ovigerous females (appendix 1). This taxonomically important structure (Defaye et al., 2000) is free distally and articulates with the body wall along a more or less marked anterior hinge.
The operculum delimits a small underlying cavity called genital atrium. The atrium opens to the exterior via the atrial slit. The genital atrium is the site of laying and fertilization of the oocytes and of the formation and attachment of the ovisac (Cuoc et al., 1989a, b; Barthélémy et al., 1998; Defaye et al., 2000). The spermatophores are deposited on the genital area, where they form seminal pseudo-receptacle, and act as an external chamber, storing sperm until fertilization (Cuoc et al., 1989a, b;
Barthélémy et al., 1998; Defaye et al., 2000). Females usually detach spent spermatophores with their modified fifth legs before extrusion of egg sacs. The intervals between the two events usually differ within and between species. For example, at an average temperature close to that in situ (3±1°C; 18:6 h light: dark cycle), females of E. graciloides detach their spermatophores in about 28 hours after mating (SD ± 21, n=25) and produce egg sac approximately 47 hours (SD ± 32, n=25) after mating. These intervals might be shorter for other species (Berger and Maier, 2001). The extruded eggs are contained into a sac and carried for some days until the nauplii hatch (Chow-Fraser & Maly, 1988; Berger and Maier, 2001). Diaptomid females lack seminal receptacles to store sperm (Cuoc et al., 1989a, b; Defaye et al., 2000) and hence re-mating is necessary to produce a second clutch of fertile eggs (Watras & Haney, 1980; Watras, 1983; Chow-Fraser & Maly, 1988; Berger and Maier, 2001).
Mating behaviour
A high proportion of diaptomid females mate multiply, as indicated by the number of externally attached spermatophores (e.g. Berger and Gerhard, 2001; Paper I, II, and
Figure 2. Scanning electron micrograph of genital area of female Eudiaptomus graciloides. Note the genital operculum (Op) and its hing (H). Scale bar: 10 µm
Figure 3. Scanning electron micrograph of spermatophore (sp) placement in Eudiaptomus graciloides. Note the genital area and placement site of the spermatophore and cement like material (right arrow). Scale bar: 10 µm
Figure 4. Photographs of spermatophores detached with fine needles from the genital area of female Eudiaptomus graciloides. A. Note the adhesive substance (right arrow) attaching all spermatophores together. Left arrow indicates spermatophore neck. B.
The two top spermatophores detached with fine needles from the adhesive substance.
Note the difference in length.
IV). Moreover, females’ rate of mating varies between species, between populations within species, and between individuals within populations. However, a single mating before each clutch production is also sufficient and does not limit female reproductive success (Watras & Haney, 1980; Watras, 1983; Chow-Fraser & Maly, 1988; Berger and Maier, 2001; Paper IV). The question of adaptive significance of polyandry in
diaptomid copepod, however, has been given much less attention despite their being an ideal organism for the topic.
Copulation involves a sequence of behavioural as well as morphological interactions between the sexes. Males actively search and chase receptive females and attempt to grasp them. Females respond to this with a vigorous escape reaction and try to dislodge the males (Watras, 1983). During this premating struggle, the male attempts to secure first a furcal and then genital double somite grasp of the female, with its right antennule and right fifth leg, respectively (Berger & Maier, 2001). These two points of attachment are critical for controlling the escaping females and for attachment of the spermatophore externally on the genital double-somite of the females. For females, body size and antennules are probably important in the behavioural response to male mating attempt as these traits are known to be directly associated with mobility and escape ability (Mauchline, 1998). Moreover, morphological segments or contact areas on females, such as the furca and genital double somite, where males make initial contact and position themselves during mating, may also be important in mediating female response to male manipulation.
Figure 5. Photograph of posterior part of female Eudiaptomus graciloides. Empty spermatophore with fertilization tube (arrow) attached to the genital opening.
Figure 6. Photograph of Eudiaptomus graciloides. Empty spermatophores with fertilization tube attached to the genital opening.
Figure 7. Photograph of posterior part of female Eudiaptomus graciloides.
Spermatophores with fertilization tube attached to the genital opening.
3. GENERAL MATERIAL AND METHODS
The studies focusing on polyandry, comprising the main part of the thesis (Paper II, III, & IV), were carried out on copepods from lake Takvatn. In Paper I, copepods from Fjellfrøsvatn were also included to provide comparative material on phenology of life cycle and reproduction, and on life history traits. The first two studies (I and II) were comparative and observational, based on fixed material collected in the field, whereas the last two studies (III & IV) were experimental studies performed on live animals under laboratory conditions with controlled light-temperature room (3±1°C;
18: 6 h light: dark cycle, provided by 20 watt incandescent lamps) and provision of lab cultured alga Scenedesmus gracilis.
The design of a sampling programme depends on the biological question, the type of information required (qualitative vs. quantitative) and the characteristics of the environment in which the studied organism lives, such as temperature, depth or predation gradients (Mauchline, 1998). In the high latitude lakes studied water temperature is low for most part of the season and copepods exhibit slowed growth rate, extended development time and seasonally restricted breeding periods (Dahl- Hansen et al., 1994; Primicerio, 2000; Klemetsen et al., 2003). Also, in the pelagic, deep area of the lakes, inhabited by E. graciloides, horizontal heterogeneity in the relevant demographic and environmental variables is low relative to temporal and between lakes heterogeneity. Therefore, given these conditions and taking into consideration the experience of previous studies (Primicerio & Klemetsen, 1999;
Primicerio, 2000), all sampling activities were carried out at fixed stations in the pelagic area, with higher sampling frequencies during the warm, open water period.
To check for spatial variation in relevant variables within lakes, additional samples were collected in 2007. Copepod samples were collected with plankton net (mesh size, 50 µm) hauled vertically at a constant speed (0.5 m s-1) from 30 m depth to the surface.
For study I, triplicate samples of copepods and water temperature were collected monthly from January 2003 to January 2004 (except in June, July and October when samples were taken twice a month) from Takvatn Lake and Fjellfrøsvatn Lake.
Additionally, triplicate samples from three different stations (>100m apart) were collected once a month from February to April in 2007. The samples were first
narcotized (5% ethanol) and then, fixed in (4%, final concentration) formalin solution for later examination. Different developmental stages and sexes were identified based on morphological characteristics (Dussart & Defaye, 1995) and counted from sub- samples under a dissection microscope. Females carrying spermatophores or eggs were also registered, and the number of spermatophores or eggs carried was counted.
Adult female body size (prosome length) from selected sampling date (that had ovigarous females) was measured with an ocular micrometer under a dissecting microscope.
For study II, I used five replicate samples collected and fixed (see above) from Takvatn Lake in February and March 2003. From each sampling date, I randomly selected females carrying from one to five spermatophores (10 females in each group) and female carrying from three to eight eggs in egg sac (10 females in each group).
Using compound microscope fitted with a drawing mirror, the image of the different body parts, spermatophores and eggs were taken onto a piece of paper and measurements of the drawings were later made to the nearest 0.01 mm. Natural female mating rate and clutch size of females was assessed, by counting the number of externally attached spermatophores and the number of eggs carried in the sac, respectively. Selection of morphological traits for investigation was made based on their importance during mating. Measured traits were analysed to examine their relationship with mating frequency and clutch size.
Mate choice experiment (III) was carried out using live copepods collected from Takvatn in January 2004. The samples were diluted in lake water in 30 l plastic tanks and transported to controlled light-temperature room. Adult males and adult unmated females were sorted and kept separately, to prevent them from mating before the experiment started. They were fed with the alga S. gracilis at the final concentration of about 5x104 cells ml-1. After four days of separation, 3 randomly selected receptive females were combined with an adult male in a small vessel containing filtered lake water. A total of 150 experimental trials were performed. After an introduction period of 24 hours, copepods were preserved for subsequent morphometric analyses (see above). The presence of a spermatophore or a fertilized egg on the female genital double somite was considered as evidence of mating. In 58 of my experimental units, I found one mated female and two unmated females. However, as some animals were
lost during handling for size measurement, I ended up with 39 experimental units for statistical analysis, each having one male, one mated and two unmated females.
Another experimental study (IV) was performed under laboratory conditions using wild-mated females that had been collected and brought to the laboratory in March 2007. In controlled light-temperature room, I randomly selected 150-mated female, 75 of which were single mated and the rest (75) double-mated, and kept the separately in groups according to their spermatophore (s) number. After 3 days of separation, egg- producing females were transferred to individual chambers containing 20 ml of filtered lake water. As diptomid females need re-mating to produce a new set of fertile eggs (e.g., Berger & Maier, 2001), I introduced a male to each chamber on the same day. The male remained present until the death of the female; if the male died earlier, it was replaced by a new male. During the 3 days period after separation, a total of 62 females (26 single mated and 36 double mated) produced egg sacs, 10 female died (4 single mated and 6 double mated) and the rest (78) failed to produce eggs.
Animals in the individual chambers were fed with the alga S. gracilis at the final concentration of about 5x104 cells ml-1 three times per week. Females were examined daily for survival and other fitness parameters. Their clutch size and the number of fertilized and unfertilized eggs were noted. Moreover, the date of hatching, the number of egg hatched and the number of alive and dead offspring were recorded.
Additionally, females were checked for any change in reproductive phases, mating and additional clutch production. The picture of live adult females and hatching were taken using a light microscope mounted with a digital camera. Measures of morphological data from the digital images of animals were later obtained using CorelDraw 11 and Canvas 8 digital image analysis software. All trait measurements were made to the nearest 0.001 mm. The experiment lasted for two months and provided data about several components of female fitness that was compared between single mated females and double-mated females.
Preliminary studies were also carried out to gather information that were relevant for the experimental studies (III & IV), i.e., time to mating and the length of time between: (i) mating (spermatophore attachment) and detachment of spermatophores, (ii) mating and egg sac production and (iii) egg production and hatching of offspring,
(iv) the first and the second spermatophore produced by males. A total of 60 experimental trials were performed, each involving the combination of 3 randomly selected receptive females and an adult male in a small glass vessel containing filtered lake water. The experiments were carried out in the abovementioned controlled light- temperature rooms, and animals were fed as mentioned above.
Samples for the different morphological figures presented in the thesis were collected from Takvatn. The techniques employed to gather such figures include, scanning electron microscopy (SEM), transmission electron microscopy (TEM), and light microscopy.
Scanning electron microscopy (SEM). Image of the external morphology of females genital area and the images for spermatophore attachment on the genital opening were made with the SEM technique. In preparation for SEM (Dussart & Defaye, 1995), specimens were cleaned, by rinsing in an aqueous solution of potassium hydroxide.
Thereafter, the specimens were fixed, using formalin, and then dehydrated in graded ethanol baths: 50%, 70%, 90%, 95%, and 100%. After that, the specimens were transferred to acetone and critical point dried. Dried samples were mounted with double sided adhesive tape on stubs in the desired orientation under dissecting microscope and coated with gold. Observation and photographing were carried out under scanning electron microscope (model: Jeol JSM 6300).
Transmission electron microscopy (TEM). This technique was used to follow whether or not multiple spermatophores attached on the female genital field reach the genital atrium, where fertilization of ova occurs. Females with multiple spermatophores were first anaesthetized and then, fixed with 3% glutaraldehyde, buffered (pH 3, 0.1 M) mixture of sodium cacodylate solution for 40 minute at 20ºC. Thereafter, specimens were post-fixed in a solution of phosphate buffer (0.1 M) and 1% osmium tetroxide at room temperature for 30 minutes. Then, they were dehydrated in an ethanol, followed by immersion in propylene oxide and finally embedded in araldite. First, several semithine (1 µm thick) section cut were made with glass knives on an ultramicrotome and stained with toluidine blue and examined under light microscopy. The sections were made starting from external ventral face of genital area where males attach their spermatophores. Then, ultrathin sections were cut with a diamond knife and collected
on 200-mesh copper grids coated with pioloform support films. They were stained with uranyl acetate followed by lead citrate and examined and photographed with electron microscope (model: Jeol JEM 1010). Staring from the point of attachment of spermatophores on the genital field of females, several ultrathin sections were cut in sequences and examined and photographed until reaching the genital atrium of the females.
4 RESULTS AND DISCUSSION
4.1 Life cycle and natural mating frequency
As background knowledge for subsequent studies (paper II, III & IV), I investigated timing of certain critical life cycle events and reproductive strategies of populations of E. graciloides living in two neighboring lakes, Takvatn and Fjellfrøsvatn (paper I).
Analyses of the results from one year of regular monthly sampling provided valuable information on seasonal changes in proportion of different developmental stages, incidence of ovigerous and mated females, natural mating frequency of female (measured as number of spermatophores per mated female) body size and clutch size.
This information, together with the record of seasonal temperature changes, helped to define the breeding season, the developmental period, the natural mating frequency of female and the annual number of generations of the two populations.
Some aspects of the life cycle followed similar seasonal trends in the two lakes. For example, both populations have only one generation per year. Moreover, they started egg production in mid-winter under the ice cover, before any phytoplankton growth has started, but synchronized the appearance of their later copepodid stages with the seasonal availability of food. The first adults appeared in early autumn, forming the over-wintering population. These findings are in line with those of other studies reporting egg production by E. graciloides in late winter, when food availability is low (Ekman, 1964; Menne & Seitz, 1992). Copepods are able to reproduce in periods of low food availability by channeling all the resources available from lipids stored during the productive season into reproduction and maintenance of vital functions (Menne & Seitz, 1992). Eudiaptomus graciloides is also able to use low concentrations of food efficiently (Menne & Seitz, 1992). In contrast to adult copepods, which are adapted to fast for long periods, the copepodid stages of E.
graciloides need sufficient food resources to develop as well as to store energy in the form of lipids to be used for overwintering and reproduction (Menne & Seitz, 1992).
Therefore, E. graciloides from the two lakes seem to have adapted to a seasonally variable environment by careful timing of their vulnerable developmental stages to the productive season. This is one of the predicted adaptations assumed for the success of an organism in theoretical studies (Gotthard, 2001, 2004).
Moreover, the field study (I) revealed the occurrence of a moderate level of multiple mating by female E. graciloides, that varies between populations as well as between individuals within a population. On average females from Takvatn showed higher mating rate (mean ± SE: 2.4 ± 0.42) relative to females from Fjellfrøsvatn (mean ± SE: 1.7 ± 0.21). The higher predation density in Fjellfrøsvatn compared to Takvatn (Klemetsen et al., 2003) might cause a reduction in the activity of both females and males. This results in a decreased encounter rate between males and females, which tends to reduce mating frequencies (Rowe et al., 1994). The seasonal trends in multiple mating and the trends in egg-sac carrying frequency showed close correspondence and reached the seasonal peak at the same time in both lakes (i.e., March in Takvatn and April and early June in Fjellfrøsvatn). However, both mating rate and egg-sac carrying frequency showed no relationship to adult sex ratio, which, with a few exceptions, remained unchanged in favour of the females. This suggests that a skewed sex ratio does not explain the multiple mating, and that males have the potential to produce more than one spermatophore; a phenomenon confirmed under laboratory conditions. Moreover, According to Maly (1970), if predators are responsible for skewed adult sex ratios, there should be a correlation between predator density and the adult sex ratio of the two populations living in similar habitats.
Therefore, the lack of difference in sex ratio between the two lakes differing in predation density (Klemetsen et al., 2003), suggest that predation is unlikely to be responsible for the skewed sex ratio observed within each population.
4.2 Development time, body size and clutch size
Despite the similar seasonal trends, the populations also showed noteworthy differences in other important life history characteristics (paper I). For example, adult female body size and mean clutch size were always larger in Takvatn than in Fjellfrøsvatn. Moreover, in Takvatn, major egg production and development of most
early stages occurred in the cooler months, and development from egg to adult took longer time than in Fjellfrøsvatn. These findings are related to one another. For example, in many organisms clutch size is a direct function of body size. Positive relationships between female size and clutch size have been shown in many cases (e.g. Smyly, 1968; Maly, 1973,1983; Checkley, 1980; paper II & IV). Moreover, development time and body size at maturity are the two most important life history traits that are usually inversely related in most ectotherms (Stearns, 1992; Roff, 2002) including copepod (Mauchline, 1998). Additionally, temperature experienced during development is one of the environmental variables that affect the relationship between development time and final body size. For example, the majority of ectotherms grow slower but mature at a larger body size in colder environments (Atkinson 1994;
Berrigan & Charnov 1994; Sibly & Atkinson 1994). Therefore, the observed difference in clutch size between the lakes may be explained by the difference in body size between populations. The difference in body size, in turn, may be explained by a negative relationship between size at maturity and temperature experienced during development.
Yet, other important factors that might cause the observed difference in development time and final body size, and hence clutch size, are food quality and availability, and the effect of size-selective predation. The effect of food does not seem important to bring about the observed pronounced difference in the life history traits. This is because the lakes are nearly similar in morphometric, trophic, physico-chemical characteristics; in annual temperature fluctuation and in ice and snow conditions (Dahl-Hansen et al., 1994; Primicerio, 2000; Klemetsen et al., 2003; see above). The effect of selective predation may, on the other hand, deserve attention. It is well known that planktivorous Arctic char is more abundant in Fjellfrøsvatn than in Takvatn, suggesting a greater predation risk for Eudiaptomus in the former lake (Dahl-Hansen et al., 1994; Klemetsen et al., 2003; Knudsen et al., 2007). For example, when predation from Planktivorous fish that prefer large prey is important, selection may favor a shorter growth period resulting in reduced age at maturity and smaller body size. As body size and clutch size are closely related, selective pressure for smaller size may also result in smaller clutch size (Nylin & Gotthard, 1998).
Additionally, in animals that carry their eggs until hatching, for example, in copepods and cladocerans, the risk of predation by visually-oriented planktivorous fishes
increase with clutch size as increased clutch size increases the conspicuousness of prey (Svensson, 1995, 1997). Consequently, smaller body size may occur as a correlated response to selection against large clutches. Even if no conclusive answer can be reached, the effect of selective predation might best explain the observed difference in life history traits between the two lakes.
Moreover, Arctic char displays seasonality in abundance and activity (Dahl-Hansen et al., 1994; Klemetsen et al., 2003). For example, in Takvatn and Fjellfrøsvatn, earlier studies showed that feeding rate of char increased markedly in late spring, peaking in summer, and then decreasing again in autumn and remaining low during most of the winter, (with a minimum in March), followed by an increase in May (Dahl-Hansen et al., 1994; Klemetsen et al., 2003). Thus, females in Fjellfrøsvatn, which showed peak egg production in late spring and early summer, should be more exposed to predation than females in Takvatn, which showed high egg production in winter, with a peak in March. Greater exposure to visual predators may also help explain the lower mating frequencies observed in Fjellfrøsvatn relative to Takvatn.
4. 3 Benefits of polyandry
Understanding the adaptive significance of polyandry is the focus of much current research work in sexual selection and several hypotheses have been proposed to explain the phenomenon (see section 1.2.2). Under laboratory conditions, I examined the benefits of polyandry by following and comparing egg production, clutch size, fertility, hatching rate of eggs, hatchling body size and early hatchling survival from females that had been mated once or twice. I found no evidence for direct fitness benefits of polyandry in female E. graciloides. Double-mated females did not differ from single mated females in several components of female fitness measures. Yet, body size of females explained much of the variation in clutch size among females regardless of whether females had been mated once or twice.
Among the hypotheses proposed to explain the benefits of polyandry, direct benefits are thought to be the origin of polyandry. Fertilization insurance is one such female benefit promoting multiple mating. This may occur when females fail to store sufficient sperm or when male ejaculates are insufficient to fertilize the whole clutch.
Under such conditions, multiple mating might therefore enhance female fitness
simply by making ample amount of sperm available to fertilize the eggs she produce (Arnqvist & Nilsson, 2000; Hosken & Stockley, 2003). I found, however, no evidences for fertility benefits promoting polyandry in female E. graciloides. That is, the fertility and hatching rate of single mated females did not differ from that of double-mated females, suggesting that single mated females were not sperm limited.
Moreover, it has also been argued that, by mating with more than one male, females may enhance their fitness via reception of ejaculatory nutrients as shown in mating systems where males transfer spermatophores that contains both sperm and nutritious accessory gland product (Thornhill & Alcock, 1983; Pitnick et al., 1997; Arnqvist &
Nilsson, 2000; Stjernholm & Karlsson, 2000; Hosken & Stockley, 2003). During mating, male E. graciloides transfer spermatophores that contain both spermatozoa and associated seminal secretions (Blades &Youngbluth. 1980; Dussart & Defaye.
1995). However, the lack of association between spermatophore numbers and clutch size suggests that the accessory seminal secretions of spermatophores are unlikely to have a nutritional function. Therefore, reproduction in female E. graciloides does not seem to be constrained by substances derived from spermatophores increaseing the number of eggs produced. Moreover, the accessory substances of ejaculates have also been suggested to have other numerous complex effects on female reproductive performance. One of these is stimulation of egg production. Additionally, the act of mating itself have been suggested to have positive effects on female fitness as it may directly trigger female egg production (Thornhill & Alcock, 1983; Arnqvist &
Nilsson, 2000; Hosken & Stockley, 2003). According to these hypotheses, double- mated females should have higher probabilities of egg production than single-mated females. I found, however, no significant difference between the two groups of females in egg production. Therefore, a single mating before production of each clutch seems sufficient and having only one spermatophore seem not to limit female’s probability of producing egg sacs (Watras & Haney, 1980; Watras, 1983; Chow- Fraser & Maly, 1988; Berger and Maier, 2001). Additionally, there were no differences in hatchling body size and early survival of offspring from single and double-mated females. Empirical evidence from other copepods show that offspring size is proportional to the probability of survival to adulthood (e.g., Cooney & Gehrs, 1980; Wyngaard, 1986; Guisande et al., 1996) and my findings suggest no indirect benefits in terms of increased offspring viability for double-mated females. Therefore, although every possible benefit of polyandry to female E. graciloides from Takvan is
impossible to rule out, the obvious direct and indirect benefits were not evident in this study
4. 4 Morphological correlates of mating status explaining multiple mating
Given that mating is a costly activity (Wilcox, 1984; Bailey et al., 1993; Cordts &
Parttridge, 1996; Clutton-Brock & Langley, 1997; Watson et al., 1998; Arnqvist, 1989; Fairbairn, 1993; Magnhagen, 1991; Rowe, 1994; Ward, 198; Daly, 1978; Hurst et al., 1995) from which females receive no apparent benefit (paper IV), why then do females E. graciloides mate multiply? I examined the relationship between female morphology and mating frequency in wild caught female (paper II & IV).
Additionally, I investigated the relationship between female morphology and male mating success using mate choice experiments (paper III). I found that the body size (paper III) and antennules (paper II, III & IV) of females were inversely associated with female mating status (paper II) and female mating frequency (paper II & IV).
Moreover, the length of the female furca (paper and II, III & IV) and the size of genital double somite (paper II & III) were directly correlated with female mating frequency and male mating success. In both studies, the selection of morphological traits for investigations was based on their importance during mating. So, consideration of behavioural components and inter-digitations of morphological traits of the sexes before copulation (pre-contact) and during copulation (post-contact) helped to explain the pattern of nonrandom mating observed (Gauld, 1957; Watras, 1983; Roff, 1972; Blades, 1977; Maier; 1995). The patterns were consistent with the predictions and assumptions of the male manipulation hypotheses (Arnqvist, 1989b;
Rowe et al., 1994; Allen & Simmons, 1996; Arnqvist, 1997; Holland & Rice, 1998;
Blanckenhorn et al., 2000; Gavrilets et al., 2001). These results will be discussed below after considering the possible costs of mating and costs of male harassment to female copepods.
Most of the general costs of mating mentioned in the introduction are likely to apply in diaptomid. For example, pairs in copulation are, compared to single individuals, more conspicuous to visual predators (Hairston et al., 1983; Winfield & Townsend, 1983) and have reduced vigilance and escape abilities (Maier et al., 2000). During copulation, the male is attached to the female with his fifth legs. In some cases, such copulatory position can be maintained for long periods, especially at low temperatures
(Burghard & Maier, 2000). Thus, a receptive female encountering a number of males will experience increased predation risk from multiple mating. Moreover, mating females most likely have higher metabolic expenditures when transporting attached males and probably also suffers a reduced foraging efficiency compared with single females. Additionally, bodily contact with multiple partners can increase the chance of disease transmission (Hurst et al., 1995). Eudiaptomus graciloides often suffers from fungus infections, specially on the egg sac (personal observations). Fungus might be transmitted to the females during mating, and infections might affect female body condition and reproduction. Given these costs of mating, from which females receive no appearent benefit, females are expected to evolve resistance to male harassment so as to reduce the cost of superfluous mating. (Parker, 1979; Arnqvist &
Nilsson 2000; Gavrilets et al., 2001). Consistent with these predictions, diaptomid females employ resistance to male harassment in the form of escape reactions (Gauld, 1957; Watras, 1983; Roff, 1972; Blades, 1977; Maier; 1995). However, resistance or escape reactions are also costly behaviours to females in terms of increased energy consumption (Watson et al., 1998), reduced foraging time (Wilcox, 1984) and increased predation risk (Rowe, 1994). For example, antennules are traits that power the escape reaction of females. They help generate 20 times faster speed than normal speed, but using them demands a very high energy output (Mauchline, 1998).
Moreover, in copepods, swimming and feeding are inseparable (Dussart & Defaye.
1995). Therefore, male harassment, which may cause a switch from maintenance swimming in females to costly displacement swimming, disrupts female feeding (Dussart & Defaye. 1995; Mauchline, 1998). Thus, both multiple mating and escape from mating, in the face of male harassment, are costly for females. Under such condition, females simply may make the "best of a bad job” by going for the less costly option, variously termed “convenience polyandry” (Thornhill & Alcock 1983;
Arnqvist, 1989, 1992; Rowe, 1992; Rowe et al., 1994; Watson et al., 1998;
Weigensberg & Fairbairn, 1996), “female reluctance” (Rowe et al., 1994;
Blankenhorn et al., 2000) or “male manipulation” (Holland & Rice, 1998; Watson et al., 1998) models.
Models of male manipulation predict that the perceived harassment rate may vary among females due to their variability in behavioural and morphological traits.
Therefore, under any given average harassment rate, such variation will generate
